Phosphorus-31 NMR and mass spectrometry of atropinesterase and

Mar 22, 1985 - and subtilisin A (Sub) and in -chymotrypsinogen (Chymogen) was labeled with a .... sites of the same enzymes and of -chymotrypsinogen, ...
0 downloads 0 Views 1MB Size
Biochemistry 1985, 24, 6894-6903

6894

31PNMR and Mass Spectrometry of Atropinesterase and Some Serine Proteases Phosphorylated with a Transition-State Analogue A. C. M. van der Drift,*,* H . C. Beck,%W. H. Dekker,%A. G. Hulst,%and E. R. J. Wild Medical Biological Laboratory TNO and Prim Maurits Laboratory TNO, 2280 A A Rijswijk, The Netherlands Received March 22, 1985

ABSTRACT: The serine residue in the active center of atropinesterase (AtrE), a-chymotrypsin (Chymo),

and subtilisin A (Sub) and in a-chymotrypsinogen (Chymogen) was labeled with a diisopropylphosphoryl (DP) group. The labeled proteins were studied in buffered aqueous solution under various native and denaturing conditions with 31PN M R before and after being subjected to “ageing”, a process leading to conversion of the DP group into a monoisopropylphosphoryl (MP) group. Besides, the model compounds Gly-Ser(DP), Gly-Glu-Ser(DP)-Gly-OEt,and diisopropyl hydrogen phosphate were investigated under similar conditions and in other solvents with different hydrogen-bonding capacity. Mass spectrometry was used to analyze products resulting from ageing in the presence of H2180. The 31Pchemical shift of the D P proteins increases according to a simple titration curve upon lowering the pH from 9.0 to 5.0. This is ascribed to protonation of a particular histidine residue in the active center that interacts with a nearby isopropoxy group by hydrogen bonding with the ester oxygen. In DP-AtrE, hydrogen bonding at the phosphoryl oxygen dominates the interaction between substituent and protein; in the other DP proteins, nonbonding interactions become more dominant in the order Chymogen < Chymo < Sub. DP-AtrE, DP-Chymo, and DP-Sub age according to first-order kinetics. The pH dependence of the reaction rate constant k, indicates that ageing is catalyzed by the protonated histidine, which is responsible for the increase in chemical shift. The direct interaction between the phosphoryl group and the histidine is lost upon ageing whereas there is an increase in the nonbonding interaction of the remaining isopropyl group with the protein in the order Chymo < Sub < AtrE. The maximum value of k, when the histidine is fully protonated (kam)increases in the same order. Ageing of the DP enzymes occurs exclusively by C - 0 fission, yielding 2-propanol and propene. Since the amount of 2-propanol decreased and that of propene increased in the order Chymo to Sub to AtrE, the increase in k,” has been interpreted as a shift in character of ageing from mainly SN2 for Chymo to considerably SN1 for AtrE and Sub. This has been attributed to preferential stabilization of the SNl transition state by an interplay of hydrogen-bonding and nonbonding interactions between the phosphoryl group and the protein. The results indicate that the active center region in AtrE resembles that of serine proteases with regard to the so-called charge-relay system, the oxyanion hole, and the nonpolar binding crevice, suggesting that esterolysis by AtrE proceeds according to a mechanism similar to that of proteolysis by Chymo or Sub.

S e r i n e proteases and serine esterases are characterized by an extremely nucleophilic serine residue at the active site. A generally accepted criterion for identification of these enzymes is the inhibition by diisopropyl phosphorofluoridate (DFP)’ (Hartley, 1960), which results from covalent binding of a diisopropylphosphoryl (DP) group to the oxygen atom in the side chain of this serine residue (Cohen et al., 1955; Oosterbaan et al., 1955). Such a phosphorus-containing substituent opens the possibility of using 31PNMR spectrometry to study these serine hydrolases (Gorenstein & Findlay, 1976; Reeck et al., 1977; Markley, 1979). Since the magnetic properties of a 31Pnucleus depend on its electronic configuration as determined by the substituents and their bonding and nonbonding interactions with the immediate environment (Letcher & Van Wazer, 1967; Ionin, 1968; Mavel, 1973), the DP group may be a suitable label to compare the structure of the active site region in different serine hydrolases. Investigations on DP-serine proteases [cf. Markley (1 979) and Steitz & Shulman (1982)j have already shown that the chemical shift of the 31Pnucleus is a sensitive criterion for the detection of variations in the interaction between the DP group and its environment, but the data are still insufficient to allow a detailed correlation between chemical shift and the corresponding interactions. *Medical Biological Laboratory T N O . Prins Maurits Laboratory TNO.

*

Further, the catalytic activity of serine hydrolases toward hydrolysis of peptide or ester bonds is based on their ability to act as a mold for binding and stabilizing the so-called tetrahedral intermediate (Kraut, 1977). In this transition state of the enzyme-substrate complex, the carbonyl carbon of the scissile bond in the substrate has a tetrahedral configuration and is covalently linked to the 0’ of the active serine, while the carbonyl oxygen resides in the “oxyanion hole” on the enzyme surface, stabilized by hydrogen bonding with the protein. Many inhibitors of serine hydrolases also form tetrahedral adducts, which, presumably, closely resemble this transient intermediate. The DP group is considered to be such a transition-state analogue (Stroud et al., 1974; Kraut, 1977), and thus, the magnetic properties of the 31P nucleus may reflect interactions that stabilize the transition state of the enzymesubstrate complex. Finally, phosphorylated serine hydrolases in solution may be converted into a nonreactivatable form by a process called ageing [cf. Usdin (1970)l. As was shown for a-chymotrypsin and butyrylcholinesterase,ageing of DP enzymes involves the



Abbreviations: AtrE, atropinesterase; Chymo, a-chymotrypsin; Chymogen, a-chymotrypsinogen; Sub, subtilisin A; DFP, diisopropyl phosphorofluoridate; DP, (i-PrO),P(=O)- (called diisopropylphosphoryl in this paper); MP, (i-PrO)OP(=O)- (called monoisopropylphosphoryl in this paper); Tris-HCI, tris(hydroxymethy1)aminomethane hydrochloride.

0 1985 American Chemical Societv

31P N M R AND MASS SPECTROMETRY OF ATROPINESTERASE

conversion of the tertiary phosphate ester into a secondary ester, i.e., the DP group loses one isopropyl group and becomes a monoisopropylphosphoryl (MP) group (Berends, 1964). It has been suggested that elements of the enzyme structure responsible for the catalytic activity play an essential role in ageing (Berends, 1964). Because of its negative charge and its geometry, the MP group may be regarded as an even better transition-state analogue than the DP group (Kossiakoff & Spencer, 1981). Alterations in the interactions with the protein resulting from such a change in transition-state analogue will affect the magnetic properties of the 31Pnucleus, and therefore, investigation of this conversion by 31PNMR may give additional information about the transition state and the structural elements involved in ageing. Besides, investigation of the ageing mechanism, e.g., by mass spectrometric characterization of the formed products, may supplement this structural information. Previously, the active site region of atropinesterase from Pseudomonas putida strain L of the biotype A (strain PMBL-l),* a pure serine esterase, was compared with that of the serine proteases a-chymotrypsin and subtilisin A with respect to inhibition kinetics and by fluorescence and ESR spectrometry (Van der Drift et al., 1981, 1985; Van der Drift, 1983). These investigations gave strong indications that some of the structural elements forming the basis of the catalytic activity of the proteases, Le., a nonpolar primary substrate binding site and a particular histidine as part of a so-called charge-relay system, are also present in atropinesterase. The present study concerns a comparison of the DP-labeled active sites of the same enzymes and of a-chymotrypsinogen, the inactive precursor of a-chymotrypsin, by means of 31PNMR and mass spectrometry with special attention to ageing. It supplements the above-mentioned investigations. A more detailed account has been given elsewhere (Van der Drift, 1983). MATERIALS AND METHODS Proteins and Model Compounds. Salt-free a-chymotrypsin (Chymo; EC 3.4.21.1) from bovine pancreas (3 times crystallized and lyophilized; lots 58C-8 135, 40F-805 1, and 49C8015) and salt-free a-chymotrypsinogen A (Chymogen) from bovine pancreas (6 times crystallized; type 11; lot 29C-8010) were obtained from Sigma Chemical Co.; dialyzed and lyophilized subtilisin A (Sub; EC 3.4.21.14) was from NOVO Industri AIS (batches 70-4,73-1, A 8003-75, A 9001-75, and A 9005-75). Atropinesterase (AtrE) was obtained from Pseudomonas putida PMBL-1 according to the procedure of Rorsch et al. (1971) as modified by R. A. Oosterbaan et al. (personal communication). Characterization of the enzymes and determination of enzyme activities were performed as described elsewhere (Van der Drift, 1983). The phosphorylated model peptides Gly-Ser(DP) and Gly-Glu-Ser(DP)-Gly-OEt were kindly provided by Dr. H. Kienhuis (Prins Maurits Laboratory TNO); diisopropyl hydrogen phosphate and diisopropyl phosphorofluoridate (DFP) were obtained from C. de Borst (Prins Maurits Laboratory TNO). Chemicals. All chemicals were reagent-grade. Organic solvents (Merck) were distilled shortly before use. Deuterium oxide (D20; 99.8% isotopically pure; Uvasol, Merck), H2180 (9 1% isotopically pure, Ventron), citric acid (Merck), sodium citrate (Merck), Tris (J. T. Baker Chemicals B.V.), urea (Union Chimique Belge, s.a.), sodium dodecyl sulfate (SDS;

* This notation refers to the culture collection of the Medical Biological Laboratory TNO.

VOL. 24, NO. 24, 1985

6895

specially purified for biochemical work, British Drug House), guanidine hydrochloride (Gdn-HC1; Fluka, AG), and 1,4dithiothreitol (Merck) were used without further purification. Buffers for calibration of pH meters were obtained from Radiometer A I S and Beckman. Labeling with a DP Group. Chymo, Chymogen, or Sub (30-90 mg) was dissolved in 1 mL of 0.1 M Tris-HC1, pH 7.6 (protein concentration 1-3 mM). AtrE solutions of similar concentration were prepared at 4 "C from stock solutions of 0.01-0.10 mM active enzyme by concentration with Amicon magnetically stirred cells (Models 202 and 52) equipped with a PM-10 filter, followed by a further concentration with Minicon macrosolute concentrators (type A-25, Amicon). To achieve complete inhibition (>99.9%), 20 pL of 0.5 M DFP in dry, peroxide-free 2-propanol (isopropyl alcohol) was added (final DFP concentration 10 mM), and the mixture was incubated at room temperature for about 6 h or, in the case of AtrE and Chymogen, for 20 h. Nonspecific labeling was not observed. Sometimes, e.g., at high concentrations of AtrE, a slight precipitate was formed. After inhibition, this was removed by centrifugation. The remaining DFP and its hydrolysis product were removed by dialysis at 4 "C against 140 mL of buffer for at least 36 h, under continuous stirring and frequent renewal. Finally, the sample was dialyzed against 12 mL of buffer containing about 20% D 2 0 as internal standard for field-frequency locking. Buffers employed were 0.1 M Tris-HC1, pH 27.6, and 0.1 M citric acidsodium citrate, pH